Hemodialysis patients depend on arteriovenous (AV) shunts for angioaccess, usually achieved by surgically implanting vascular grafts composed of expanded polytetrafluroethylene (ePTFE). Unfortunately, approximately half of these AV grafts (AVGs) fail within 1 year due to incremental stenosis arising from vascular lesion formation (VLF). This complex process involves 1) thrombin generation, 2) platelet activation with secretion of platelet-derived growth factor (PDGF) and thrombus formation, 3) monocyte/macrophage accumulation and secretion of PDGF, and 4) intimal migration and proliferation of vascular smooth muscle cells. Since dietary n-3FAs down-regulate many of the molecular and cellular interactions underlying the development of VLF, we propose to test the hypothesis that dietary n-3FAs will reduce stenosing thrombo- occlusive graft failure. We have persuasive evidence in baboons that dietary n-3FAs interrupt thrombosis and VLF without impairing hemostatic function. We also have direct experimental evidence in dialysis patients implicating thrombosis in the processes of AVG stenosis and thrombo-occlusion, and experience in sickle cell patients receiving dietary n-3FAs demonstrating feasibility, acceptance and reduction in pain episodes. We propose to: 1. Establish the effective antithrombotic dose of dietary n-3FAs. During the first year of the project the effects of dietary n-3FAs at 0.1 vs 0.3 g/kg/d will be compared with control olive oil in 24 patients randomly allocated to three 8-patient cohorts, measuring surrogate molecular measures of thrombosis, inflammation and risk factors before and after 3 months of therapy. Differences in platelet deposition on AVGs will be compared using autologous platelets labeled with 111In-oxine and quantitative gamma camera imaging. 2. Evaluate the effects of dietary n-FAs on AVG failure in a randomized, blinded, olive off-controlled clinical trial. During the remaining three years of the project, we propose to carry out a randomized, olive oil-controlled, blinded trial in 180 dialysis patients comparing the effects of dietary n-3FAs vs equivalent control olive oil on AVG failure. The dose of dietary n-3FAs will be either 0.1 g/kg/d or 0.3 g/kg/d, depending on their relative efficacy for interrupting AVG platelet deposition. The primary outcome will be AVG failure. Assuming AVG failure rate of 50 percent per year, treatment benefit of 50 percent, and 25 percent patient loss per year, randomizing 180 dialysis patients will have 90 percent power of detecting a benefit. Secondary outcomes to be compared are molecular markers for thrombosis, inflammation, and vascular risk factors.